This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMP117: TMP117

Part Number: TMP117

Hi, 

Can you please help answer these questions from my customer? Thanks!

(1) Could the device operate safely at the lowest voltage (i.e., 1.8V +/- 0.1V or 1.9V +/- 0.1V)?

(2) What is the accuracy between 20C and 45C? Are all units tested for accuracy specs during manufacturing? Is the accuracy specification in terms of RMS (i.e., 68% confidence) or 95% confidence?

(3) Is this device recommended for measuring skin temperature in clinical settings? What is the standard used for testing the claimed accuracy?

(4) In terms of performance and power consumption, is the smallest package (1 x 1.5 mm) equivalent to the 2 x 2 mm?

(5) If a temperature reading is taken every 1 second, what is the optimal configuration (i.e., i2c clock frequency and other settings) that maintains claimed accuracy and minimizes power consumption? What would be the total power consumption for this configuration? Will it change significantly if the supply voltage varies from 3.6V to 1.8V?

(6) Does the claimed accuracy change significantly with supply voltage?

(7) If we decide not to use the alert pin, should we leave it open or grounded/vcc with a resistor? Does it impact power consumption?

  • HI Soumya,

    Thanks for your questions!

    Soumya said:
    (1) Could the device operate safely at the lowest voltage (i.e., 1.8V +/- 0.1V or 1.9V +/- 0.1V)?

    The TMP117 will not work correctly with a rail below 1.8 V, but will work fine as long as that minimum threshold is met. so 1.9 +/- 0.1 V would work, but not 1.8 +/- 0.1 V

    Soumya said:
    (2) What is the accuracy between 20C and 45C? Are all units tested for accuracy specs during manufacturing? Is the accuracy specification in terms of RMS (i.e., 68% confidence) or 95% confidence?

    I'm checking on this for you, I will try to respond today or Monday with a solid answer.

    Soumya said:
    (3) Is this device recommended for measuring skin temperature in clinical settings? What is the standard used for testing the claimed accuracy?

    Yes, this is the device we recommend for this purpose. You can use the top-side of the package, or the bottom side of the board with vias to touch a metal contact made of some bio compatible material such as stainless steel. That contact can conduct heat to the device from the skin. 

    Soumya said:
    (4) In terms of performance and power consumption, is the smallest package (1 x 1.5 mm) equivalent to the 2 x 2 mm?

    Both devices will meet all the same performance specs. For certain applications (such as measuring temperature using the top of the device) the BGA package would be recommended. Also consider that the smaller package will more quickly respond, as it has a smaller thermal mass. Power consumption between the two options will be the same.

    Soumya said:
    (5) If a temperature reading is taken every 1 second, what is the optimal configuration (i.e., i2c clock frequency and other settings) that maintains claimed accuracy and minimizes power consumption? What would be the total power consumption for this configuration?

    As long as you meet the conditions in the accuracy table of the datasheet you will achieve the recommended accuracy. That said, here are my recommended settings:

    • 1 Hz Conversion rate
    • 8 Averages
    • 400kHz I2C, taking reads around once a second
    • Use I2C Pullups that are roughly 5kOhm

    With these settings I measured just below 50uA average consumption in the lab on one of our EVMs. 

    Ideally you would put the device in a shutdown state between reads, and trigger a one-shot measurement once a second from your controller, but it is up to your customer if that is worth it. It would reduce the current consumption by probably a few uA. 

    Soumya said:
    Will it change significantly if the supply voltage varies from 3.6V to 1.8V?

    Current consumption is about the same from 1.8 to 5.5V. It does increase slightly with higher voltage but not enough for any heating to be an issue. Lower voltage will generally mean better accuracy, but the accuracy spec will hold all the way to 5.5V. 

    Soumya said:
    (6) Does the claimed accuracy change significantly with supply voltage?

    Not significantly, we hold the accuracy spec from 1.8 V to 5.5 V, but operating at lower voltages does improve the available margin.

    Soumya said:
    (7) If we decide not to use the alert pin, should we leave it open or grounded/vcc with a resistor? Does it impact power consumption?

    It can be left floating or grounded, both are acceptable. If the alert pin is not being used it will not affect power consumption. 

    I hope this information is helpful. I will check on the accuracy form 20 to 45 C, but if you have any additional questions in the meantime please feel free to ask. 

    Best Regards,
    Brandon Fisher

  • Hi Soumya,

    Following up regarding your accuracy question. We can't provide a spec for the 20 to 45C region of the TMP117M, so it would be considered +/- 0.2C as shown in the datasheet.

    The TMP117 and TMP117N would both be +/- 0.1C in that range.

    Best Regards,
    Brandon Fisher